We investigate the finite source process of the October 16, 1999 Hector Mine earthquake. The earthquake ruptured nearly 40 km of the Bullion fault (Southern California) which is located 20 km east-north-east of the Landers - Emerson - Camp Rock fault system which ruptured in the Landers Mw7.3 earthquake on June 28, 1992 (Figure 16.1). The proximity in both space and time of the large Hector Mine earthquake to the earlier Landers earthquake is very intriguing, suggesting that there is a relationship between the two. In the region of the Hector Mine hypocenter the Coulomb stress change on the Bullion fault due to the Landers earthquake transitions from less favorable to more favorable for right-lateral slip on the northwest trending fault (Parsons and Dreger, 2000).
Broadband displacement data at seven stations (Figure 16.1) at regional and local distances are used to determine the slip distribution in space and time for a 3-segment, multiple time window fault model. The inversion is formulated as a least square problem subject to slip positivity and smoothing constraints. We also use a surface slip constraint (Scientists from USGS, SCEC, CDMG, 2000) on the slip values of the upper row of subfaults. The Green's functions were calculated for the SoCal model using frequency-wavenumber integration method (Saikia, 1994).
The obtained slip distribution is shown on the Figure 16.2. The largest asperity is confined to the Lavic Lake fault and has maximum slip equal to 640 cm; slip amplitude on the Bullion fault is nearly half as large. When the slip on the overlapping portions of the fault is summed (Figure 16.2d), the highest amplitudes appear below the hypocenter. The earthquake appears to be essentially bilateral and covers an approximately 35 by 15 km area. The maximum overall slip is 823 cm and the total moment obtained in the inversion is 8.2e19 Nm. This value is compatible to other studies that take the finite source process into account, i.e. 6.1e19 Nm (Yagi and Kikuchi, 1999), 6.0e19 Nm (Ji et al., 2000), 6.2e19Nm (E. Price, pers. comm., 2000). The fault length is almost twice smaller than could be anticipated from the empirical relationship for events of this magnitude derived by Wells and Coppersmith (1994). This discrepancy is justified through comparison with surface slip and by the long duration of the event.
The time history of the event (not shown) shows that within first 3 time windows (around 3 seconds) there are dispersed patches of slip with relatively low amplitude. The slip starts appearing in the fourth window that begins 3.6 seconds after rupture initiation. The maximum in slip falls between 7.2 and 9.6 seconds after the origin time and it then decays steadily in subsequent windows. In all windows with appreciable slip, most of it is located in the vicinity of the hypocenter. The moment rate function, obtained by integration of the moment distribution in the space domain, demonstrates that the slow initiation of the event lasts about 3-5 seconds. The major part of the moment is released within the first 20 seconds, which is a relatively long time for the source dimensions. Due to the complex time history of the event, we are not able to determine the best constant rupture velocity. We prefer the value of 2.2 km/s, because inspection of the time windows reveals a delay in rupture that is consistent with the velocity and acceleration seismograms.
In order to investigate a possibility of a dip-slip component, we solved the inverse problem simultaneously for a rake of -180 and 90 degrees allowing us to resolve variable rake. The results indicate that the amplitude of dip-slip is much smaller than that of the strike-slip and only a 2% improvement in variance reduction is obtained in this case. The distribution of aftershocks from the CNSS catalog shows that to the north of the hypocenter most of the events occurred to the east of the Lavic Lake fault. It may indicate that there was some coseismic rupture, or possibly this fault represents post-seismic activity. To check these hypotheses, we perform an inversion for a fault model, which includes the northern extension of the middle segment B (see Figure 16.1), fixing all other inversion parameters equal to those of the preferred model. Overall we find that the resulting slip is very similar to the preferred model (Figure 16.2). To the north of the hypocenter, most of the slip occurred on the Lavic Lake fault. From our analysis there does not seem to be any requirement by the data for appreciable slip on the fault east of the Lavic Lake fault in contrast to the results of Ji et al. (2000). The aftershocks along this trace may be due to co-seismic stress adjustments, or perhaps triggered by low levels of sympathetic slip.
The source slip model may be used to simulate near-fault strong ground motions. We have been developing a system to use near-regional distance displacement seismograms to constrain the seismic moment tensor, the best nodal plane, and the slip distribution and to use this information to produce a strong ground motion map (e.g., ShakeMap). The October 16, 1999 Hector Mine earthquake provided an opportunity to test those procedures (Dreger and Kaverina, 2000). We have demonstrated that it is possible to predict near-fault strong shaking from the derived finite fault parameters. The determination of the best fault plane and slip distribution were obtained the morning of the earthquake, and were subsequently verified by the observed fault offsets and aftershock locations. The predicted PGV were found to agree with the sparse near-fault observations and the TriNet ShakeMap. This method may be applied in both densely and sparsely instrumented regions. In densely instrumented regions it may provide needed redundancy in the event of telemetry or other failure of the primary realtime strong-motion network. In sparsely instrumented areas it may be the only viable approach to estimate the level of near-fault strong ground motions. Although the software was not automated at the time of the earthquake, it was possible to simulate a realtime application. The computer processing times of the individual components of the method as applied to the Hector Mine earthquake indicate that the method may be applied in near-realtime and provide strong shaking information within the 30-minute post-earthquake time frame. Current work is focussing on the implementation of this methodology into the REDI processing system.
Dreger, D. S., and A. Kaverina, Seismic remote sensing for the earthquake source process and near-source strong shaking: a case study of the October 16, 1999 Hector Mine earthquake, Geophys. Res. Lett., 27, 1941-1944, 2000.
Ji, C., Wald, D.J., and D. V. Helmberger, Slip history of 1999 Hector Mine, California earthquake, Seism. Res. Lett., 71, 224, 2000.
Parsons, T., and D. S. Dreger, Static-stress impact of the 1992 Landers earthquake sequence on nucleation and slip at the site of the 199 M=7.1 Hector Mine earthquake, southern California, Geophys. Res. Lett., 27, 1949-1952, 2000.
Scientists from USGS, SCEC, CDMG, Preliminary report on the 16 October 1999 M7.1 Hector Mine, California, earthquake, Seism. Res. Lett., 71, 11-23, 2000.
Saikia, C. K., Modified frequency-wave-number algorithm for regional seismograms using Filon's quadrature - modeling of L(g) waves in eastern North America, Geophys. J. Int., 118, 142-158, 1994.
Yagi, Y. and M. Kikuchi, Preliminary results of rupture process for the October 16, 1999 Hector Mine Earthquake, http://wwweic.eri.u-tokyo.ac.jp/yuji/Hector /Hector.html, 1999.
Wells, D. L. and K. J. Coppersmith, New empirical relationships among magnitude, rupture length, rupture width, rupture area, and surface displacement, Bull. Seism. Soc. Am., 84, 974-1002, 1994.